(164) Use of Real-World Data to Evaluate Overall Survival Following Robotic-Assisted Versus Non-Robotic Assisted Radical Prostatectomy

Background: Radical prostatectomy (RP) is the most common definitive local therapy for prostate cancer, especially for those with cancer at risk for spread. While most RPs are now performed via robotic-assisted surgical device (RASD), there is a need for evidence to evaluate long-term survival for RASD versus non-RASD approaches.

Objectives: To determine if 5- through 10-year overall survival (OS) after RASD RP is non-inferior to non-RASD RP.

Methods: This observational study selected males aged ≥35 years with treatment-naive prostate cancer with an RP procedure code between 7/1/2007 (start of available data after requiring ≥180 days of baseline enrollment) and 12/31/2014 (allowing for ≥5 years of follow-up prior to the start of COVID-19) from Optum CDM claims data. Patients were categorized by procedure code as ‘RASD’ if the code indicated robotic assistance and ‘non-RASD’ if the code indicated perineal or retropubic open RP approaches. OS was defined as the absence of a reported death, with death data sourced from linked national Death Master File data available during enrollment and after disenrollment. To report OS, patients were followed from the day after RP until the end of the follow-up period (e.g., 5 years for 5-year OS) or 12/31/2019 (end of study period). To report the difference in OS probability, a Kaplan-Meier approach was used after propensity score (PS) stratification in quintiles to control for measured confounding. To test for a non-inferiority margin of 2% with a test size of 2.5%, a hierarchical approach was used among identical populations beginning with 5-year analysis and requiring significance before proceeding to the 6-year analysis (and so on, up to 10 years), precluding the need for multiple testing adjustment.

Results: Among the patients undergoing RASD and non-RASD RP (18,949 vs. 5,401), unadjusted OS was improved at 5 (96.13% vs 95.65%) through 10 years (89.18% vs. 87.31%). Crude differences in 5- and 10-year OS probabilities per 100 patients for RASD versus non-RASD were 0.48 (95% CI: -0.13, 1.09) and 1.87 (0.75, 2.99), respectively; attenuated to 0.20 (-0.19, 0.59) and 0.88 (0.12, 1.63) after PS stratification demonstrating non-inferiority (p < 0.0001 for both).

Conclusions: Among male patients with treatment-naive prostate cancer, RP procedures performed via RASD were non-inferior to non-RASD procedures with respect to overall survival at 5 through 10 years post-RP, and suggest improved survival associated with RASD at 10 years.